Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]
Groups > linux.kernel > #1650863
| From | "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> |
|---|---|
| Newsgroups | linux.kernel |
| Subject | [PATCH tip/core/rcu 66/88] srcu: Prevent sdp->srcu_gp_seq_needed counter wrap |
| Date | 2017-05-26 00:10 +0200 |
| Message-ID | <tL97d-5CD-51@gated-at.bofh.it> (permalink) |
| References | <tL8Xv-5iU-1@gated-at.bofh.it> |
| Organization | linux.* mail to news gateway |
If a given CPU never happens to ever start an SRCU grace period, the
grace-period sequence counter might wrap. If this CPU were to decide to
finally start a grace period, the state of its sdp->srcu_gp_seq_needed
might make it appear that it has already requested this grace period,
which would prevent starting the grace period. If no other CPU ever started
a grace period again, this would look like a grace-period hang. Even
if some other CPU took pity and started the needed grace period, the
leaf rcu_node structure's ->srcu_data_have_cbs field won't have record
of the fact that this CPU has a callback pending, which would look like
a very localized grace-period hang.
This might seem very unlikely, but SRCU grace periods can take less than
a microsecond on small systems, which means that overflow can happen
in much less than an hour on a 32-bit embedded system. And embedded
systems are especially likely to have long-term idle CPUs. Therefore,
it makes sense to prevent this scenario from happening.
This commit therefore scans each srcu_data structure occasionally,
with frequency controlled by the srcutree.counter_wrap_check kernel
boot parameter. This parameter can be set to something like 255
in order to exercise the counter-wrap-prevention code.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
---
Documentation/admin-guide/kernel-parameters.txt | 9 +++++++++
kernel/rcu/srcutree.c | 18 ++++++++++++++++++
2 files changed, 27 insertions(+)
diff --git a/Documentation/admin-guide/kernel-parameters.txt b/Documentation/admin-guide/kernel-parameters.txt
index 01b5ab92d251..6671f9b60a86 100644
--- a/Documentation/admin-guide/kernel-parameters.txt
+++ b/Documentation/admin-guide/kernel-parameters.txt
@@ -3810,6 +3810,15 @@
spia_pedr=
spia_peddr=
+ srcutree.counter_wrap_check [KNL]
+ Specifies how frequently to check for
+ grace-period sequence counter wrap for the
+ srcu_data structure's ->srcu_gp_seq_needed field.
+ The greater the number of bits set in this kernel
+ parameter, the less frequently counter wrap will
+ be checked for. Note that the bottom two bits
+ are ignored.
+
srcutree.exp_holdoff [KNL]
Specifies how many nanoseconds must elapse
since the end of the last SRCU grace period for
diff --git a/kernel/rcu/srcutree.c b/kernel/rcu/srcutree.c
index 31203469b2d1..b4058d2a4e8d 100644
--- a/kernel/rcu/srcutree.c
+++ b/kernel/rcu/srcutree.c
@@ -45,6 +45,10 @@
static ulong exp_holdoff = DEFAULT_SRCU_EXP_HOLDOFF;
module_param(exp_holdoff, ulong, 0444);
+/* Overflow-check frequency. N bits roughly says every 2**N grace periods. */
+static ulong counter_wrap_check = (ULONG_MAX >> 2);
+module_param(counter_wrap_check, ulong, 0444);
+
static void srcu_invoke_callbacks(struct work_struct *work);
static void srcu_reschedule(struct srcu_struct *sp, unsigned long delay);
@@ -497,10 +501,13 @@ static void srcu_gp_end(struct srcu_struct *sp)
{
unsigned long cbdelay;
bool cbs;
+ int cpu;
+ unsigned long flags;
unsigned long gpseq;
int idx;
int idxnext;
unsigned long mask;
+ struct srcu_data *sdp;
struct srcu_node *snp;
/* Prevent more than one additional grace period. */
@@ -539,6 +546,17 @@ static void srcu_gp_end(struct srcu_struct *sp)
smp_mb(); /* GP end before CB invocation. */
srcu_schedule_cbs_snp(sp, snp, mask, cbdelay);
}
+
+ /* Occasionally prevent srcu_data counter wrap. */
+ if (!(gpseq & counter_wrap_check))
+ for (cpu = snp->grplo; cpu <= snp->grphi; cpu++) {
+ sdp = per_cpu_ptr(sp->sda, cpu);
+ spin_lock_irqsave(&sdp->lock, flags);
+ if (ULONG_CMP_GE(gpseq,
+ sdp->srcu_gp_seq_needed + 100))
+ sdp->srcu_gp_seq_needed = gpseq;
+ spin_unlock_irqrestore(&sdp->lock, flags);
+ }
}
/* Callback initiation done, allow grace periods after next. */
--
2.5.2
Back to linux.kernel | Previous | Next — Previous in thread | Next in thread | Find similar | Unroll thread
[PATCH tip/core/rcu 0/88] Commits for 4.13 "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:00 +0200
[PATCH tip/core/rcu 55/88] rcu: Eliminate the unused __rcu_is_watching() function "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:10 +0200
[PATCH tip/core/rcu 30/88] srcu: Document auto-expediting requirement "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:10 +0200
[PATCH tip/core/rcu 49/88] rcu: Move docbook comments out of rcupdate.h "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:10 +0200
[PATCH tip/core/rcu 60/88] rcu: Move rcu_ftrace_dump() from rcupdate.h to rcu.h "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:10 +0200
[PATCH tip/core/rcu 57/88] rcu: Remove linux/debugobjects.h from rcupdate.h "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:10 +0200
[PATCH tip/core/rcu 34/88] rcu: Update rcu_bootup_announce_oddness() "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:10 +0200
[PATCH tip/core/rcu 13/88] rcu: Remove obsolete reference to synchronize_kernel() "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:10 +0200
[PATCH tip/core/rcu 08/88] rcutorture: Update test scenarios based on new Kconfig dependencies "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:10 +0200
[PATCH tip/core/rcu 54/88] rcu: Move rcupdate.h to new empty-function style "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:10 +0200
[PATCH tip/core/rcu 24/88] rcutorture: Reduce CPUs dedicated to testing Classic SRCU "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:10 +0200
[PATCH tip/core/rcu 77/88] rcu: Remove #ifdef moving rcu_end_inkernel_boot from rcupdate.h "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:10 +0200
[PATCH tip/core/rcu 78/88] rcu: Remove typecheck() from RCU locking wrapper functions "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:10 +0200
[PATCH tip/core/rcu 70/88] rcu: Refactor #includes from include/linux/rcupdate.h "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:10 +0200
[PATCH tip/core/rcu 36/88] srcu: Print non-default exp_holdoff values at boot time "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:10 +0200
[PATCH tip/core/rcu 01/88] rcutorture: Add lockdep to one of the SRCU scenarios "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:10 +0200
[PATCH tip/core/rcu 51/88] rcu: Move expediting-related access/control out of rcupdate.h "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:10 +0200
[PATCH tip/core/rcu 61/88] rcu: move rcupdate.h to the new true/false-function style "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:10 +0200
[PATCH tip/core/rcu 66/88] srcu: Prevent sdp->srcu_gp_seq_needed counter wrap "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:10 +0200
[PATCH tip/core/rcu 71/88] rcu: Convert rnp->lock wrappers to macros for SRCU use "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:10 +0200
[PATCH tip/core/rcu 52/88] rcu: Move torture-related definitions from rcupdate.h to rcu.h "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:10 +0200
[PATCH tip/core/rcu 58/88] rcu: Improve __call_rcu() debug-objects error message "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:10 +0200
[PATCH tip/core/rcu 73/88] srcu: Use rnp->lock wrappers to replace explicit memory barriers "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:10 +0200
[PATCH tip/core/rcu 45/88] rcu: Add memory barriers for NOCB leader wakeup "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:10 +0200
[PATCH tip/core/rcu 62/88] rcu: Move torture-related functions out of rcutiny.h and rcutree.h "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:10 +0200
[PATCH tip/core/rcu 23/88] srcu: Make Classic and Tree SRCU announce themselves at bootup "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:20 +0200
[PATCH tip/core/rcu 38/88] rcu: Add lockdep_assert_held() teeth to tree_plugin.h "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:20 +0200
[PATCH tip/core/rcu 43/88] sched: Rely on synchronize_rcu_mult() de-duplication "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:20 +0200
[PATCH tip/core/rcu 67/88] srcu: Shrink srcu.h by moving docbook and private function "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:20 +0200
[PATCH tip/core/rcu 56/88] rcu: Move the RCU_SCHEDULER_ definitions from rcupdate.h "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:20 +0200
[PATCH tip/core/rcu 32/88] rcu: Add preemptibility checks in rcu_sched_qs() and rcu_bh_qs() "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:20 +0200
[PATCH tip/core/rcu 16/88] rcu: Make sync_rcu_preempt_exp_done() return bool "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:20 +0200
[PATCH tip/core/rcu 39/88] srcu: Make SRCU be once again optional "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:20 +0200
[PATCH tip/core/rcu 26/88] rcuperf: Set more user-friendly defaults "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:20 +0200
[PATCH tip/core/rcu 10/88] rcu: Complain if blocking in preemptible RCU read-side critical section "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:20 +0200
[PATCH tip/core/rcu 84/88] rcu: Eliminate NOCBs CPU-state Kconfig options "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:20 +0200
[PATCH tip/core/rcu 04/88] rcutorture: Add a scenario for Tiny SRCU "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:20 +0200
[PATCH tip/core/rcu 80/88] rcu: Remove SPARSE_RCU_POINTER Kconfig option "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:20 +0200
[PATCH tip/core/rcu 12/88] rcuperf: Remove conflicting Kconfig options "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:20 +0200
[PATCH tip/core/rcu 47/88] rcu: Flag need for rcu_node_tree.h and rcu_segcblist.h visibility "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:20 +0200
[PATCH tip/core/rcu 79/88] rcu: Remove the now-obsolete PROVE_RCU_REPEATEDLY Kconfig option "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:20 +0200
[PATCH tip/core/rcu 25/88] srcu: Shrink Tiny SRCU a bit more "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:20 +0200
[PATCH tip/core/rcu 75/88] rcu: Remove the RCU_KTHREAD_PRIO Kconfig option "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:20 +0200
[PATCH tip/core/rcu 85/88] rcu: Move RCU non-debug Kconfig options to kernel/rcu "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:20 +0200
[PATCH tip/core/rcu 65/88] rcu: move rcutiny.h to the new true/false-function style "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:20 +0200
[PATCH tip/core/rcu 74/88] rcu: Remove *_SLOW_* Kconfig options "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:20 +0200
[PATCH tip/core/rcu 86/88] rcu: Move RCU debug Kconfig options to kernel/rcu "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:20 +0200
[PATCH tip/core/rcu 63/88] rcu: Move rcu_request_urgent_qs_task() out of rcutiny.h and rcutree.h "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:20 +0200
[PATCH tip/core/rcu 42/88] rcu: Make synchronize_rcu_mult() check for duplicates "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:20 +0200
[PATCH tip/core/rcu 48/88] rcutorture: Don't wait for kernel when all builds fail "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:20 +0200
[PATCH tip/core/rcu 82/88] srcu: Remove Classic SRCU "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:20 +0200
[PATCH tip/core/rcu 76/88] rcu: Remove nohz_full full-system-idle state machine "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:20 +0200
Re: [PATCH tip/core/rcu 76/88] rcu: Remove nohz_full full-system-idle state machine Linus Torvalds <torvalds@linux-foundation.org> - 2017-05-26 01:10 +0200
Re: [PATCH tip/core/rcu 76/88] rcu: Remove nohz_full full-system-idle state machine "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 01:50 +0200
[PATCH tip/core/rcu 07/88] rcutorture: Correctly handle CONFIG_RCU_TORTURE_TEST_* options "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:20 +0200
[PATCH tip/core/rcu 40/88] srcu: Shrink Tiny SRCU a bit "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:20 +0200
[PATCH tip/core/rcu 50/88] rcu: Move rcu_expedited and rcu_normal externs from rcupdate.h "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:20 +0200
[PATCH tip/core/rcu 87/88] rcu: Remove event tracing from Tiny RCU "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:20 +0200
[PATCH tip/core/rcu 53/88] rcu: Remove UINT_CMP_GE() and UINT_CMP_LT() "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:20 +0200
[PATCH tip/core/rcu 27/88] srcu-cbmc: Use /usr/bin/awk instead of /bin/awk "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:20 +0200
[PATCH tip/core/rcu 72/88] rcu: Move rnp->lock wrappers for SRCU use "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:20 +0200
[PATCH tip/core/rcu 44/88] rcu: Use RCU_NOCB_WAKE rather than RCU_NOGP_WAKE "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:20 +0200
[PATCH tip/core/rcu 68/88] srcu: Apply trivial callback lists to shrink Tiny SRCU "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:20 +0200
[PATCH tip/core/rcu 41/88] srcu: Add DEBUG_OBJECTS_RCU_HEAD functionality "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:20 +0200
[PATCH tip/core/rcu 28/88] rcuperf: Add writer_holdoff boot parameter "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:20 +0200
[PATCH tip/core/rcu 59/88] rcu: Move rcu_is_nocb_cpu() from rcupdate.h to rcu.h "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:20 +0200
[PATCH tip/core/rcu 69/88] lockdep: Use consistent printing primitives "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:20 +0200
[PATCH tip/core/rcu 31/88] doc: Take tail recursion into account in RCU requirements "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:20 +0200
[PATCH tip/core/rcu 46/88] torture: Add --kconfig argument to kvm.sh "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:20 +0200
[PATCH tip/core/rcu 18/88] rcuperf: Add test for dynamically initialized srcu_struct "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:20 +0200
[PATCH tip/core/rcu 22/88] rcuperf: Add the ability to test tiny RCU flavors "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:20 +0200
[PATCH tip/core/rcu 02/88] rcutorture: Add three-level tree test for Tree SRCU "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:20 +0200
[PATCH tip/core/rcu 88/88] rcu: Remove RCU CPU stall warnings from Tiny RCU "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:20 +0200
[PATCH tip/core/rcu 03/88] rcutorture: Fix bug in reporting Kconfig mis-settings "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:20 +0200
[PATCH tip/core/rcu 06/88] rcu: Prevent rcu_barrier() from starting needless grace periods "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:20 +0200
[PATCH tip/core/rcu 37/88] rcu: Add lockdep_assert_held() teeth to tree.c "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:20 +0200
[PATCH tip/core/rcu 64/88] rcu: Move rcutiny.h to new empty-function style "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:20 +0200
[PATCH tip/core/rcu 81/88] srcu: Fix rcutorture-statistics typo "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:20 +0200
[PATCH tip/core/rcu 20/88] atomics: Add header comment so spin_unlock_wait() "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:20 +0200
[PATCH tip/core/rcu 19/88] doc/atomic_ops: Clarify smp_mb__{before,after}_atomic() "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:30 +0200
[PATCH tip/core/rcu 21/88] docs: Fix typo in Documentation/memory-barriers.txt "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:30 +0200
[PATCH tip/core/rcu 15/88] rcuperf: Add a Kconfig-fragment file for Classic SRCU "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:30 +0200
[PATCH tip/core/rcu 35/88] srcu: Make exp_holdoff module parameter be static "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:30 +0200
[PATCH tip/core/rcu 09/88] srcu: Eliminate possibility of destructive counter overflow "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:30 +0200
[PATCH tip/core/rcu 29/88] rcutorture: Add "git diff" output to testid.txt file "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:30 +0200
[PATCH tip/core/rcu 11/88] rcuperf: Defer expedited/normal check to end of test "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:30 +0200
[PATCH tip/core/rcu 17/88] checkpatch: Remove checks for expedited grace periods "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:30 +0200
[PATCH tip/core/rcu 05/88] rcutorture: Add a scenario for Classic SRCU "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:30 +0200
[PATCH tip/core/rcu 14/88] rcuperf: Add ability to performance-test call_rcu() and friends "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:30 +0200
[PATCH tip/core/rcu 33/88] rcu: Print out rcupdate.c non-default boot-time settings "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> - 2017-05-26 00:30 +0200
csiph-web